Administration

Organization, billing and SSO

Plans and seats, the trial, single sign-on, workspace policy, health checks, and self-hosted updates.

10 min read

A workspace is one organization: its people, its suppliers, its campaigns, its plan. This article covers what an admin sets once and rarely touches again, plus the two operational screens on a self-hosted install.

Plans and seats

Plan	What it is
Free trial	14 days, 3 seats, from self-serve sign-up. Everything works.
Standard	The self-serve paid plan. Bought with a card from the Billing screen.
Business	Adds single sign-on. Sales-led.
Enterprise	Sales-led.

A seat is an internal account: a Campaign Manager or an Admin, counted role-blind. Supplier contacts are unlimited and free, however many campaigns they are on and however many tags they hold. Going over your seat count warns rather than blocks, and seats reconcile at renewal.

The trial

A trial workspace shows a countdown banner under the top bar. What happens if it runs out is stated everywhere it is relevant, because a surprise is the wrong thing to spring on somebody who has put a quarter of promotions into a product.

Day 0: sign-up
14 days, 3 seats, whoever signed up is the workspace admin.
Day 14: the trial ends
The workspace goes read-only and its campaigns are archived. Nothing is deleted. You can still read and export everything.
Day 53: a final warning
The admins are emailed that deletion is a week away.
Day 60: deletion
The workspace and its data are permanently deleted: campaigns, submissions, uploaded files, and any accounts whose only tie was this workspace.

Billing

Billing, admins only. It shows your current plan, your seat usage, and the trial countdown if you are on one.

Upgrade
Upgrade opens a hosted checkout page. Paying converts the workspace to Standard, unarchives anything the trial archived, and lifts the read-only lock.
Manage an existing subscription
Manage billing opens the payment provider's own customer portal: card details, invoices, cancellation. There is deliberately no second billing UI inside the product to keep out of step with it.
Need more than Standard
Business and Enterprise are sales-led. The screen tells you how to reach us.

Single sign-on

SSO, admins on the Business plan. Your team signs in through Microsoft Entra ID, Okta, Google, or any OIDC provider. Supplier contacts are unaffected and keep their own accounts.

Register the redirect URI with your provider
The screen prints the exact URI to paste, ending in /api/auth/sso/callback.
Enter the connection details
Issuer URL, Client ID, Client secret, and an email domain. The issuer takes a provider-specific form: Entra wants https://login.microsoftonline.com/<tenant-id>/v2.0, Okta wants https://<org>.okta.com, Google wants https://accounts.google.com.
Save, then test from a private window
Use the Use single sign-on link on the sign-in page. Testing in a private window means a broken configuration cannot lock out the session you would use to fix it.

Rule	Detail
Domain gating	Only verified email addresses on the domain you named can sign in through the connection.
Just-in-time accounts	A first sign-in creates a Campaign Manager seat automatically.
Admin stays manual	SSO never grants admin. Promote people on the Users screen.
Passwords still work	Disabling SSO returns your team to passwords and the reset flow. Nobody is stranded.
Deactivation still works	A deactivated account cannot sign in, by SSO or by password.

Personal settings

Your nameSettings, for everybody including suppliers.

Setting	Effect
Name	How you appear in the audit trail and to your colleagues.
Timezone	Every deadline and timestamp you see anywhere, including in emails, is rendered in it.
Date format	Four choices: `MM/DD/YYYY`, `DD/MM/YYYY`, `YYYY-MM-DD`, `MMM D, YYYY`. Also applies to emails.
Password	Minimum 8 characters. Replaced by a Sign-in section on an SSO-managed account.

Workspace policy

Workspace holds the settings that apply to everybody in the organization. Admins only.

Item comments

A comment thread on a single line, between you and the supplier contacts who hold that line's tag. Two controls, because not every organization wants free-form conversation kept in a purchasing tool:

Control	What it does
Allow comment threads	Off removes threads from every screen and refuses new ones, including from a tab that was already open. Comments already written are kept rather than deleted, so switching it back on restores them.
Keep comments for N days	The nightly sweep permanently deletes comments older than this. Leave it blank to keep them, which is what the audit trail and the email log do. The thread says the window above the box, so people know before they write.

Health and updates

Two screens for a self-hosted install. On a managed deployment they are informational.

Health

Health is a list of checks, each green, amber or red, with a plain-language detail line. A summary at the top says either that everything is healthy or exactly how many things need attention.

Check	What it answers
Virus scanning	Whether uploaded files are being scanned, and by what.
File backups	Whether every uploaded file is being copied to a second bucket. Reported separately from database backups because the two fail separately, and one green line covering both would hide whichever one broke.
Database backups	When the last one was taken, how many are kept, and the tool that took it. If backups are not possible at all on this deployment, for example because the database tools are not installed, the check says that rather than reporting a missing file.
Point-in-time recovery	Whether the database is recording enough to be restored to a chosen moment rather than only to last night's backup. Watch this one: it fails quietly, and the first sign is usually somebody needing a restore that is not there.
Database connections	How many of the server's connections are in use, and the cap this instance is allowed. Running out looks like the app getting slower rather than like anything with a name, so it is worth watching before you add a second instance.
Storage capacity	Where uploaded files are kept and whether that place can run out. On a bucket it cannot, so the check reports how many files are stored; on a server disk it reports the space left.
Outbound email	Sends that failed in the last week, and sends that were skipped because no mail key is configured.
Outbound queue	Messages waiting, in flight, and given up on. A queue that stopped draining shows as an old oldest-waiting rather than a large count, so the age is reported and not only the number.
Daily summaries	Updates being held for people who asked for one email a day, and how long the oldest has waited. Anything past about two days means a summary stopped being built, and somebody is hearing nothing.
Slow work	Anything that took longer than the slow threshold in the last day. See What has been slow below.
Last update	The version running, and whether the last update succeeded.
Pending invitations	Supplier invitations that have not been emailed yet.

What has been slow

Below the checks, when there is anything to show, is a table of the work that took longest in the last 24 hours. It is grouped: twenty samples of one statement are one problem, not twenty. Each row is either an operation, which is a step the code named, or a statement, which is database work with its values already replaced by placeholders. The columns are how many times it happened, the worst time and the average.

The point of the table is to make a problem findable while it is small. Use it to answer "what has got slower since last month" rather than to watch a number: samples are kept for two weeks and then removed.

System

System shows the running version and requests updates.

If no supervisor is running, the screen says so and the update button does nothing useful, which is the honest answer rather than a failure ten minutes later. See Troubleshooting for what to check when something looks wrong.